reef-pi :: An opensource reef tank controller based on Raspberry Pi.

Ranjib · Nov 22, 2018

wykat said:
With the PCA9685 and MCP23017 technically we could control >1'000 outputs with just a single PI. Each component/module has 16 channels.

yes

. I think we'll run our of compute power on pi, before electronics limitations, once mcp23017 support lands. We have to update the reef-pi code as well, currently it assumes only one pca9685 is present.
Whats more interesting is Raspberry Pi 4 will be based on different processor, means significantly more firepower. And god knows what else peripheral, may be more built in PWMs, or even analog I/O .. finger crossed

theatrus · Nov 23, 2018

Ranjib said:
yes . I think we'll run our of compute power on pi, before electronics limitations, once mcp23017 support lands. We have to update the reef-pi code as well, currently it assumes only one pca9685 is present.
Whats more interesting is Raspberry Pi 4 will be based on different processor, means significantly more firepower. And god knows what else peripheral, may be more built in PWMs, or even analog I/O .. finger crossed

Finding analog I/O or PWM in smartphone SoCs would be very rare - it’s not going to be in the core CPU. It could pivot to an embedded SoC like the Beagleboard family with its MCU co-processors, or just tack on an analog I/O front end. I’d imagine RPi would stay with Broadcom for the corporate support (and consuming all the excess inventory

).

It would probably be best to standardize on an MCU for General I/O and analog, and just use it as a programmable I/O hub sort of like a data acquisition and control board. Most microcontrollers will have oodles of analog inputs, PWMs, and pins. This would necessitate some pre-programming of the part, but some have USB or I2C bootloaders from the factory letting ReefPi manage firmware.

Ranjib · Nov 23, 2018

theatrus said:
Finding analog I/O or PWM in smartphone SoCs would be very rare - it’s not going to be in the core CPU. It could pivot to an embedded SoC like the Beagleboard family with its MCU co-processors, or just tack on an analog I/O front end. I’d imagine RPi would stay with Broadcom for the corporate support (and consuming all the excess inventory ).

It would probably be best to standardize on an MCU for General I/O and analog, and just use it as a programmable I/O hub sort of like a data acquisition and control board. Most microcontrollers will have oodles of analog inputs, PWMs, and pins. This would necessitate some pre-programming of the part, but some have USB or I2C bootloaders from the factory letting ReefPi manage firmware.

I’m going to think about this . After the baseboard and ph board is out for general consumption I’ll have a better idea on what it would take to support reef-pi with its specific hardware dependencies

Nicolas Sabrun · Nov 23, 2018

I look at the light controller guide. It specified the use of truelumen strip as a moon light.

I have a couple of spare truelumen pro led. They work on 24 VDC.

Does the same concept applies? Meaning cam i use the same mosfet and a 24 v power supply to control the led. ( which come with the led). Each strip uses 9 w i believe and i would like to use 3 of them.

Is this possible?

Also, these leds are a bit strong for moonlight so i would like to be able to dim them or even better have them ramp up/down.

Can it be done and how would you do it?

These leds don t accept pwm. I know they can be dim with and inline dimmer on the power cords. However, I got no clue on what to use to accomplish this with reef pi.

Thank you as always for your help.

theatrus · Nov 23, 2018

Ranjib said:
I’m going to think about this . After the baseboard and ph board is out for general consumption I’ll have a better idea on what it would take to support reef-pi with its specific hardware dependencies

I’m thinking along the lines of the Atmel SAMD21 series

Why?

Cheap
Good peripherals
Not archaic
Already used by the Arduino Zero and Circuit Playground Express (you can run MicroPython on the big one!)
I was already use the SAMD10/11 on the Pico board for pH and system watchdog

It’s not a DIP package and not 5V I/O as the main downsides. But someone could reprogram a CircuitPlaygroundExpress or Arduino Zero as the controller.

Ranjib · Nov 23, 2018

theatrus said:
I’m thinking along the lines of the Atmel SAMD21 series

Why?

Cheap
Good peripherals
Not archaic
Already used by the Arduino Zero and Circuit Playground Express (you can run MicroPython on the big one!)
I was already use the SAMD10/11 on the Pico board for pH and system watchdog

It’s not a DIP package and not 5V I/O as the main downsides. But someone could reprogram a CircuitPlaygroundExpress or Arduino Zero as the controller.

Can we reuse any of the adafruit micro controller boards, like feather or trinket or metro m0 ?

Soulreefer · Nov 23, 2018

halow Ranjib, I need your help to check my schematic, sorry I am not electronic engineer maybe there are still many errors, I have a problem with libraries for CHIP L29D tipe SMD. I am very happy if anyone can help this eagle library chip. thanks

This design is not finished just to see the chip layout.

theatrus · Nov 23, 2018

Ranjib said:
Can we reuse any of the adafruit micro controller boards, like feather or trinket or metro m0 ?

Yes. Metro M0 is an Arduino Zero clone. Will have to look at the other two.

Ranjib · Nov 23, 2018

Soulreefer said:
halow Ranjib, I need your help to check my schematic, sorry I am not electronic engineer maybe there are still many errors, I have a problem with libraries for CHIP L29D tipe SMD. I am very happy if anyone can help this eagle library chip. thanks

This design is not finished just to see the chip layout.

Im not an expert on these things, but from my limited knowledge, I think the circuit is ok. MCP23017 VDD input should have a capacitor?

theatrus · Nov 23, 2018

Soulreefer said:
halow Ranjib, I need your help to check my schematic, sorry I am not electronic engineer maybe there are still many errors, I have a problem with libraries for CHIP L29D tipe SMD. I am very happy if anyone can help this eagle library chip. thanks

This design is not finished just to see the chip layout.

You’re trying using an LM2596 as a boost converter (5 to 12V?)? It’s only capable of buck operation. When designing with switching converters you’ll want to pay close attention to the layout - use the datasheet as a reference.

http://www.ti.com/lit/ds/symlink/lm2596.pdf

HM3105 · Nov 23, 2018

So, I'm having an issue with telemetry. Reef-Pi 2.0 on Pi0.

When I put the AIO key in and hit "update" it gives me an error and says "please set a valid password".

Email alerts aren't enabled.

The log says "Invalid API key provided"

pickupman66 · Nov 23, 2018

Guys, I took my DCI apart today and here are images. I ran the multimeter across them to verify resistance and r3 was 2.66k ohm. Rest are readable in the images. Lmk if this tells you guys anything about what we passed thru to the jebao wp pumps from the 0-10v analog signal.

Mwdsonny · Nov 23, 2018

Any chance you can set it to calibrate nema motors? like 100 steps per mm. then you can say does 5mm and it will tell the motor to move 500 steps?

Ranjib · Nov 23, 2018

Mwdsonny said:
Any chance you can set it to calibrate nema motors? like 100 steps per mm. then you can say does 5mm and it will tell the motor to move 500 steps?

currently, the calibration features only allow speed (percentage of a single PWM channel) and duration.

Ranjib · Nov 23, 2018

pickupman66 said:
Guys, I took my DCI apart today and here are images. I ran the multimeter across them to verify resistance and r3 was 2.66k ohm. Rest are readable in the images. Lmk if this tells you guys anything about what we passed thru to the jebao wp pumps from the 0-10v analog signal.

Theres nothing on the opposite side of the pcb ?

Ranjib · Nov 23, 2018

HM3105 said:
So, I'm having an issue with telemetry. Reef-Pi 2.0 on Pi0.

When I put the AIO key in and hit "update" it gives me an error and says "please set a valid password".

Email alerts aren't enabled.

The log says "Invalid API key provided"

Can be a software bug, let me test this out and get back to you

Ranjib · Nov 23, 2018

Ranjib said:
Can be a software bug, let me test this out and get back to you

Its a bug :-( . https://github.com/reef-pi/reef-pi/issues/651 , I'll fix it soon.
Meanwhile, you should be able to save the telemetry config by adding bogus (non-empty) values in username/password section of email setting, just keep it disabled and reef-pi wont send any email.

pickupman66 · Nov 23, 2018

Ranjib said:
Theres nothing on the opposite side of the pcb ?

Nope. Nothing

Mwdsonny · Nov 23, 2018

Ranjib said:
currently, the calibration features only allow speed (percentage of a single PWM channel) and duration.

what would it take to get steps (number of pulses per ml) in the program? control multiple dosers? and each pump would need its own calibration cause of viscosity of different fluids.

HM3105 · Nov 23, 2018

Ranjib said:
Its a bug :-( . https://github.com/reef-pi/reef-pi/issues/651 , I'll fix it soon.
Meanwhile, you should be able to save the telemetry config by adding bogus (non-empty) values in username/password section of email setting, just keep it disabled and reef-pi wont send any email.

Ok, any ideas why it's not sending the data to adafruit though ?